Abstract. Let Pk := F2x1, x2, . . . , xk be the polynomial algebra over the prime field of two elements, F2, in k variables x1, x2, . . . , xk, each of degree 1. We study the hit problem, set up by F. Peterson, of finding a minimal set of generators for Pk as a module over the mod2 Steenrod algebra, A. In this paper, we study a minimal set of generators for Amodule Pk in some socall generic degrees and apply these results to explicitly determine the hit problem for k = 4.
ON THE PETERSON HIT PROBLEM NGUYỄN SUM Abstract Let Pk := F2 [x1 , x2 , , xk ] be the polynomial algebra over the prime field of two elements, F2 , in k variables x1 , x2 , , xk , each of degree We study the hit problem, set up by F Peterson, of finding a minimal set of generators for Pk as a module over the mod-2 Steenrod algebra, A In this paper, we study a minimal set of generators for A-module Pk in some so-call generic degrees and apply these results to explicitly determine the hit problem for k = Dedicated to Prof N H V Hưng on the occasion of his sixtieth birthday Introduction and statement of results Let Vk be an elementary abelian 2-group of rank k Denote by BVk the classifying space of Vk It may be thought of as the product of k copies of the real projective space RP∞ Then Pk := H ∗ (BVk ) ∼ = F2 [x1 , x2 , , xk ], a polynomial algebra in k variables x1 , x2 , , xk , each of degree Here the cohomology is taken with coefficients in the prime field F2 of two elements Being the cohomology of a space, Pk is a module over the mod Steenrod algebra A The action of A on Pk can explicitly be given by the formula xj , i = 0, Sq i (xj ) = x2j , i = 1, 0, otherwise, and subject to the Cartan formula n Sq n (f g) = Sq i (f )Sq n−i (g), i=0 for f, g ∈ Pk (see Steenrod and Epstein [29]) A polynomial f in Pk is called hit if it can be written as a finite sum f = i + + i>0 Sq (fi ) for some polynomials fi That means f belongs to A Pk , where A denotes the augmentation ideal in A We are interested in the hit problem, set up by F Peterson, of finding a minimal set of generators for the polynomial algebra Pk as a module over the Steenrod algebra In other words, we want to find a basis of the F2 -vector space QPk := Pk /A+ Pk = F2 ⊗A Pk The hit problem was first studied by Peterson [21, 22], Wood [36], Singer [27], and Priddy [23], who showed its relationship to several classical problems respectively in cobordism theory, modular representation theory, Adams spectral sequence for 12010 Mathematics Subject Classification Primary 55S10; 55S05, 55T15 2Keywords and phrases: Steenrod squares, polynomial algebra, Peterson hit problem NGUYỄN SUM the stable homotopy of spheres, and stable homotopy type of classifying spaces of finite groups The vector space QPk was explicitly calculated by Peterson [21] for k = 1, 2, by Kameko [14] for k = The case k = has been treated by Kameko [16] and by us [30] Several aspects of the hit problem were then investigated by many authors (See Boardman [1], Bruner, Hà and Hưng [2], Carlisle and Wood [3], Crabb and Hubbuck [4], Giambalvo and Peterson [5], Hà [6], Hưng [7], Hưng and Nam [8, 9], Hưng and Peterson [10, 11], Janfada and Wood [12, 13], Kameko [14, 15], Minami [17], Mothebe [18], Nam [19, 20], Repka and Selick [24], Singer [28], Silverman [25], Walker and Wood [33, 34, 35], Wood [37, 38] and others.) The µ-function is one of the numerical functions that have much been used in the context of the hit problem For a positive integer n, by µ(n) one means the smallest number r for which it is possible to write n = i r (2di − 1), where di > A routine computation shows that µ(n) = s if and only if there exists uniquely a sequence of integers d1 > d2 > > ds−1 ds > such that n = 2d1 + 2d2 + + 2ds−1 + 2ds − s it implies n − s is even and µ( n−s s ) by (QPk )n the subspace of QPk consisting From this Denote by homogeneous polynomials of degree n in Pk (1.1) of all the classes represented Peterson [21] made the following conjecture, which was subsequently proved by Wood [36] Theorem 1.1 (Wood [36]) If µ(n) > k, then (QPk )n = One of the main tools in the study of the hit problem is Kameko’s homomorphism Sq ∗ : QPk → QPk This homomorphism is induced by the F2 -linear map, also denoted by Sq ∗ : Pk → Pk , given by Sq ∗ (x) = y, if x = x1 x2 xk y , 0, otherwise, for any monomial x ∈ Pk Note that Sq ∗ is not an A-homomorphism However, 0 Sq ∗ Sq 2t = Sq t Sq ∗ , and Sq ∗ Sq 2t+1 = for any non-negative integer t Theorem 1.2 (Kameko [14]) Let m be a positive integer If µ(2m + k) = k, then Sq ∗ : (QPk )2m+k → (QPk )m is an isomorphism of GLk -modules Based on Theorems 1.1 and 1.2, the hit problem is reduced to the case of degree n with µ(n) = s < k The hit problem in the case of degree n of the form (1.1) with s = k − 1, di−1 − di > for i < k and dk−1 > was studied by Crabb and Hubbuck [4], Nam [19] and Repka and Selick [24] In this paper, we explicitly determine the hit problem for the case k = First, we study the hit problem for the cases of degree n of the form (1.1) for s = k − The following theorem gives an inductive formula for the dimension of (QPk )n in this case ON THE PETERSON HIT PROBLEM Theorem 1.3 Let n = i k−1 (2di − 1) with di positive integers such that d1 > d2 > > dk−2 dk−1 , and let m = i k−2 (2di −dk−1 − 1) If dk−1 k − 1, then dim(QPk )n = (2k − 1) dim(QPk−1 )m For dk−1 k, the theorem follows from the results in Nam [19] and the present author [32] However, for dk−1 = k − 1, the theorem is new Based on Theorem 1.3, we explicitly compute QP4 Theorem 1.4 Let n be an arbitrary positive integer with µ(n) < The dimension of the F2 -vector space (QP4 )n is given by the following table: n 2s+1 − 2s+1 − 2s+1 − 2s+2 + 2s+1 − 2s+3 + 2s+1 − 2s+4 + 2s+1 − 2s+t+1 + 2s+1 − 3, t 2s+1 + 2s − 2s+2 + 2s − 2s+3 + 2s − 2s+4 + 2s − 2s+5 + 2s − 2s+t + 2s − 2, t 2s+2 + 2s+1 + 2s − 2s+3 + 2s+2 + 2s − 2s+t+1 + 2s+t + 2s − 3, t 2s+3 + 2s+1 + 2s − 2s+u+1 + 2s+1 + 2s − 3, u 2s+u+2 + 2s+2 + 2s − 3, u 2s+t+u + 2s+t + 2s − 3, u 2, t s=1 s=2 s=3 s=4 s 15 35 45 45 24 50 70 80 14 35 75 89 85 46 94 105 105 105 87 135 150 150 150 136 180 195 195 195 150 195 210 210 210 21 70 116 164 175 55 126 192 240 255 73 165 241 285 300 95 179 255 300 315 115 175 255 300 315 125 175 255 300 315 64 120 120 120 120 155 210 210 210 210 140 210 210 210 210 140 225 225 225 225 120 210 210 210 210 225 315 315 315 315 210 315 315 315 315 The space QP4 was also computed in Kameko [16] by using computer calculation However the manuscript is unpublished at the time of the writing Carlisle and Wood showed in [3] that the dimension of the vector space (QPk )m is uniformly bounded by a number depended only on k In 1990, Kameko made the following conjecture in his Johns Hopkins University PhD thesis [14] Conjecture 1.5 (Kameko [14]) For every nonnegative integer m, (2i − 1) dim(QPk )m i k The conjecture was shown by Kameko himself for k in [14] From Theorem 1.4, we see that the conjecture is also true for k = By induction on k, using Theorem 1.3, we obtain the following NGUYỄN SUM Corollary 1.6 Let n = 2, di−1 − di i − 1, i di i k−1 (2 k − 1, dk−1 − 1) with di positive integers If d1 − d2 k − 1, then (2i − 1) dim(QPk )n = i k For the case di−1 − di i, i k − 1, and dk−1 k, this result is due to Nam [19] This corollary also shows that Kameko’s conjecture is true for the degree n as given in the corollary By induction on k, using Theorems 1.3, 1.4 and the fact that the dual of the Kameko squaring is an epimorphism, one gets the following Corollary 1.7 Let n = i k−2 (2di −1) with di positive integers and let dk−1 = 1, nr = i r−2 (2di −dr−1 −1)−1 with r = 5, 6, , k If d1 −d2 4, di−2 −di−1 i, for i k and k 5, then (2i − 1) + dim(QPk )n = i k (2i − 1) dim Ker(Sq ∗ )nr , r k r+1 i k (Sq ∗ )nr where : (QPr )2nr +r → (QPr )nr denotes the squaring operation Sq ∗ in degree 2nr + r Here, by convention, r+1 i k (2i − 1) = for r = k This corollary has been proved in [32] for the case di−2 − di−1 > i + with i k Obviously 2nr +r = i r−2 (2ei −1), where ei = di −dr−1 +1 for i r−2 er−2 e1 e2 So, in degree 2nr + r of Pr , there is a so-called spike x = x21 −1 x22 −1 x2r−2 −1 , i.e a monomial whose exponents are all of the form 2e − for some e Since the class [x] in (QPk )2nr +r represented by the spike x is nonzero and Sq ∗ ([x]) = 0, we have Ker(Sq ∗ )nr = 0, for any r k Therefore, by Corollary 1.7, Kameko’s conjecture is not true in degree n = 2nk + k for any k 5, where nk = 2d1 −1 + 2d2 −1 + + 2dk−2 −1 − k + This paper is organized as follows In Section 2, we recall some needed information on the admissible monomials in Pk and Singer’s criterion on the hit monomials We prove Theorem 1.3 in Section by describing a basis of (QPk )n in terms of a given basis of (QPk−1 )m In Section 4, we recall the results on the hit problem for k Theorem 1.4 will be proved in Section by explicitly determining all of the admissible monomials in P4 The first formulation of this paper was given in a 240-page preprint in 2007 [30], which was then publicized to a remarkable number of colleagues One year latter, we found the negative answer to Kameko’s conjecture on the hit problem [31, 32] Being led by the insight of this new study, we have remarkably reduced the length of the paper Preliminaries In this section, we recall some results in Kameko [14] and Singer [28] which will be used in the next sections ON THE PETERSON HIT PROBLEM Notation 2.1 Throughout the paper, we use the following notations Nk = {1, 2, , k}, XI = Xi1 ,i2 , ,ir = x1 x ˆ i1 x ˆ ir xk xi , I = {i1 , i2 , , xir } ⊂ Nk , = i∈Nk \I In particular, we have XNk = 1, X∅ = x1 x2 xk , Xi = x1 x ˆi xk , i k Let αi (a) denote the i-th coefficient in dyadic expansion of a nonnegative integer a That means a = α0 (a)20 + α1 (a)21 + α2 (a)22 + , for αi (a) = or and i Denote by α(a) the number of one in dyadic expansion of a Let x = xa1 xa2 xakk ∈ Pk Denote by νj (x) = aj , j k Set Ii (x) = {j ∈ Nk : αi (νj (x)) = 0}, for i Then we have i XI2i (x) x= i For a polynomial f in Pk , we denote by [f ] the class in QPk represented by f For a subset S ⊂ Pk , we denote [S] = {[f ] : f ∈ S} ⊂ QPk Definition 2.2 For a monomial x, define two sequences associated with x by ω(x) = (ω1 (x), ω2 (x), , ωi (x), ), σ(x) = (a1 , a2 , , ak ), where ωi (x) = j k αi−1 (νj (x)) = deg XIi−1 (x) , i The sequence ω(x) is called the weight vector of x (see Wood [37]) The weight vectors and the sigma vectors can be ordered by the left lexicographical order Let ω = (ω1 , ω2 , , ωi , ) be a sequence of nonnegative integers such that ωi = for i Define deg ω = i>0 2i−1 ωi Denote by Pk (ω) the subspace of Pk spanned by all monomials y such that deg y = deg ω, ω(y) ω and Pk− (ω) the subspace of Pk spanned by all monomials y ∈ Pk (ω) such that ω(y) < ω Denote j by A+ j < 2s s the subspace of A spanned by all Sq with Definition 2.3 Let ω be a sequence of nonnegative integers and f, g two homogeneous polynomials of the same degree in Pk i) f ≡ g if and only if f − g ∈ A+ Pk − ii) f (s,ω) g if and only if f − g ∈ A+ s Pk + Pk (ω) − Since A+ (0,ω) g if and only if f − g ∈ Pk (ω) If x is a monomial in Pk = 0, f Pk and ω = ω(x), then we denote x s g if and only if x (s,ω(x)) g Obviously, the relations ≡ and (s,ω) are equivalence relations We recall some relations on the action of the Steenrod squares on Pk Proposition 2.4 Let f be a homogeneous polynomial in Pk i) If i > deg f , then Sq i (f ) = If i = deg f , then Sq i (f ) = f s s s s ii) If i is not divisible by 2s , then Sq i (f ) = while Sq r2 (f ) = (Sq r (f ))2 NGUYỄN SUM Proposition 2.5 Let x, y be monomials and f, g homogeneous polynomials in Pk such that deg x = deg f , deg y = deg g s s i) If ωi (x) for i > s and x s f , then xy s f y s s ii) If ωi (x) = for i > s, x s f and y r g, then xy s+r f g Proof Suppose that Sq i (zi ) = h ∈ Pk− (ω(x)) x+f + i s, then ωi (x) = 1, ωi (z) = Then we have i−1 2j−1 ωj (x) = αi−1 2i−1 + αi−1 deg x − j=1 2j−1 ωj (x)) = j>i On the other hand, since deg x = deg z, ωi (z) = and ωj (z) = ωj (x), j = 1, 2, , i − 1, one gets i−1 i−1 j−1 αi−1 deg x − 2j−1 ωj (z) ωj (x) = αi−1 deg z − j=1 j=1 2j−1 ωj (z) = = αi−1 j>i This is a contradiction Hence i s From these about equalities and the fact that h ∈ Pk− (ω(x)), one gets s s s s s Sq i (zi y ) = hy ∈ Pk− (ω(xy )) xy + f y + i s2 > > sr−1 sr > and sj = for j > r, then it is called a minimal spike The following is a criterion for the hit monomials in Pk NGUYỄN SUM Theorem 2.12 (Singer [28]) Suppose x ∈ Pk is a monomial of degree n, where µ(n) k Let z be the minimal spike of degree n If ω(x) < ω(z), then x is hit From this theorem, we see that if z is a minimal spike, then Pk (ω(z)) ⊂ A+ Pk The following lemmas were proved in [32] Lemma 2.13 ([32]) Let n = i k−1 (2di − 1) with di positive integers such that d1 > d2 > > dk−2 dk−1 > 0, and x a monomial of degree n in Pk If [x] = 0, then ωi (x) = k − for i dk−1 di Lemma 2.14 ([32]) Let n = − 1) with di positive integers such i k−1 (2 that d1 > d2 > > dk−2 dk−1 > 0, and x a monomial in Pk such that ωi (x) = k − 1, for i = 1, 2, , s dk−1 and ωi (x) = for i > s Suppose y, f and g are polynomials in Pk with deg f = deg x and deg y = deg g = (n − deg x)/2s = 2d1 −s + + 2dk−2 −s + 2dk−1 −s − k + s s i) If x s f , then xg ≡ f g s s ii) If y ≡ g, then xy ≡ xg For latter use, we set Pk0 = {x = xa1 xa2 xakk ; a1 a2 ak = 0} , Pk+ = {x = xa1 xa2 xakk ; a1 a2 ak > 0} It is easy to see that Pk0 and Pk+ are the A-submodules of Pk Furthermore, we have the following Proposition 2.15 We have a direct summand decomposition of the F2 -vector spaces QPk = QPk0 ⊕ QPk+ Here QPk0 = Pk0 /A+ Pk0 and QPk+ = Pk+ /A+ Pk+ Proof of Theorem 1.3 We denote Nk = {(i; I); I = (i1 , i2 , , ir ), i < i1 < < ir k, r < k} Let (i; I) ∈ Nk and j ∈ Nk Denote by r = (I) the length of I, and I ∪j = I, (i1 , , it−1 , j, it , , ir ), if j ∈ I, if it−1 < j < it , t r + Here i0 = and ir+1 = k + For h < k, we set Nh−1 ∪ h = {(i; I ∪ h); (i; I) ∈ Nh−1 } Then we have Nk = (N1 ∪ 2) ∪ ∪ (Nk−1 ∪ k) ∪ {(1; ∅), , (k; ∅)} For i substituting (3.1) k, define the homomorphism fi = fk;i : Pk−1 → Pk of algebras by fi (xj ) = xj , xj+1 , if if i j < i, j < k ON THE PETERSON HIT PROBLEM Definition 3.1 Let (i; I) ∈ Nk , let r = (I), and let u be an integer with u r A monomial x ∈ Pk−1 is said to be u-compatible with (i; I) if all of the following hold: i) νi1 −1 (x) = νi2 −1 (x) = = νi(u−1) −1 (x) = 2r − 1, ii) νiu −1 (x) > 2r − 1, iii) αr−t (νiu −1 (x)) = 1, ∀t, t u, iv) αr−t (νit −1 (x)) = 1, ∀t, u < t r Clearly, a monomial x can be u-compatible with a given (i; I) ∈ Nk , r = (I) > 0, for at most one value of u By convention, x is 1-compatible with (i; ∅) r−1 r−u r−t Definition 3.2 Let (i; I) ∈ Nk , x(I,u) = x2iu + +2 for u u 0, then x is u-compatible with (i; I) and d d φ(i;I) (x) = φ(iu ;Ju ) (X −1 )fi (y)2 , (3.3) where Ju = (iu+1 , , ir ) Let B be a finite subset of Pk−1 consisting of some homogeneous polynomials in degree n We set Φ0 (B) = φ(i;∅) (B) = i k fi (B) i k Φ+ (B) = φ(i;I) (B) \ Pk0 (i;I)∈Nk ,0< (I) k−1 Φ(B) = Φ0 (B) Φ+ (B) It is easy to see that if Bk−1 (n) is a minimal set of generators for Pk−1 in degree n, then Φ0 (Bk−1 (n)) is a minimal set of generators for A-module Pk0 in degree n and Φ+ (Bk−1 (n)) ⊂ Pk+ Proposition 3.3 Let n = i k−1 (2di − 1) with di positive integers such that d1 > d2 > > dk−2 dk−1 k−1 If Bk−1 (n) is a minimal set of generators for A-module Pk−1 in degree n, then Bk (n) = Φ(Bk−1 (n)) is also a minimal set of generators for A-module Pk in degree n For dk−1 k, this proposition is a modification of a result in Nam [19] For dk−2 = dk−1 > k, it has been proved in [32] We prepare some lemmas for the proof of this proposition 10 NGUYỄN SUM Lemma 3.4 Let j0 , j1 , , jd−1 ∈ Nk Then there is (i; I) ∈ Nk such that t Xj2t x= d−1 φ(i;I) (X d −1 ), t d2 > > dk−2 dk−1 > 0, and let y0 be a monomial in (Pk )m−1 , yi = y0 xi for i k, and (i; I) ∈ Nk i) If < r = (I) < d = dk−1 , then φ(i;I) (X d −1 d )yi2 ≡ φ(j;I) (X d −1 d )yj2 + j k −1, then d d D = E = ∅ If dk−2 > dk−1 = k − 1, then E = ∅ We set B¯ = {¯ z ; X −1 z¯2 ∈ B} If d d either d k or I = I1 , then φ(i;I) (z) = φ(i;I) (X −1 )fi (¯ z )2 If d = dk−1 = k − 1, then d 2d −1 )f1 (¯ z )2 , if z ∈ C, φ(2;I2 ) (X d d φ(1;I1 ) (z) = φ(3;I3 ) (X −1 )f2 (¯ (3.4) z )2 , if z ∈ D, 2d −1 2d φ(4;I4 ) (X )f3 (¯ z ) , if z ∈ E For any (i; I) ∈ Nk , we define the homomorphism p(i;I) : Pk → Pk−1 of algebras by substituting if j < i, xj , p(i;I) (xj ) = x , if j = i, s∈I s−1 xj−1 , if i < j k 46 NGUYỄN SUM we see that either x71 x62 x3 x34 or x71 x52 x23 x34 is a term of C Since x71 x62 x3 x34 is not a term of C , the monomial x71 x52 x23 x34 is a term of C Then we have (Sq )3 (θ1 + Sq (x71 x62 x3 x34 + x71 x52 x23 x34 )) = (Sq )3 (Sq (C ) + Sq (D)), where C = C + x71 x52 x23 x34 = C + x71 x62 x3 x34 + x71 x52 x23 x34 Now the monomial x = x71 x12 x3 x4 is a term of (Sq )3 (θ1 + Sq (x71 x62 x3 x34 + x71 x52 x23 x34 )) Hence either x71 x62 x3 x34 or x71 x52 x23 x34 is a term of C is a term of C On the other hand, the two monomials x71 x62 x3 x34 and x71 x52 x23 x34 are not the terms of C We have a contradiction Hence one gets γ20 = Step Since γ20 = 0, the homomorphism ϕ2 sends (5.4.6.3) to γ25 [θ1 ] + γ31 [θ2 ] + γ5 [θ3 ] = (5.4.6.4) Using (5.4.6.4) and the result in Step 1, we get γ5 = Step The homomorphism ϕ3 sends (5.4.6.3) to γ25 [θ4 ] + γ31 [θ2 ] = (5.4.6.5) Using the relation (5.4.6.5) and the result in Step 2, we obtain γ25 = Step Since ϕ4 ([θ2 ]) = [θ1 ], we have γ31 [θ1 ] = Using this equality and by a same argument as given in Step 3, we get γ31 = For t > 1, we have |B4+ (n)| = m(t) with m(2) = 95, m(3) = 128 and m(t) = 139 for t Suppose there is a linear relation m(t) S= γi di ≡ 0, (5.4.6.6) i=1 with γi ∈ F2 and di = dn,i A direct computation from the relations p(j;J) (S) ≡ 0, for (j; J) ∈ N4 , we obtain γi = for all i The proposition is proved 5.5 The case of degree 2s+t + 2s − For s and t 2, the space (QP4 )n was determined in [32] Hence, in this subsection we need only to compute (QP4 )n for n = 2s+1 + 2s − with s > Recall that, the homomorphism Sq ∗ : (QP4 )2s+1 +2s −2 → (QP4 )2s +2s−1 −3 is an epimorphism Hence we have (QP4 )2m+4 ∼ = (QP4 )m ⊕ (QP40 )2m+4 ⊕ (KerSq ∗ ∩ (QP4+ )2m+4 ), where m = 2s + 2s−1 − So it suffices to compute KerSq ∗ ∩ (QP4+ )n for s > For s > 1, denote by C(s) the set of all the following monomials: s s+1 x1 x2 x23 −2 x24 −2 , s+1 s x1 x22 −2 x3 x24 −2 , s+1 s x1 x22 −1 x23 x24 −4 , s+1 s x1 x32 x23 −4 x24 −2 , s+1 s s s+1 x1 x2 x23 −2 x24 −2 , x1 x22 −2 x3 x24 −2 , s s+1 s+1 s x1 x22 x23 −4 x24 −1 , x1 x22 x23 −1 x24 −4 , s+1 s s+1 s x21 −1 x2 x23 x24 −4 , x1 x22 x32 −3 x24 −2 , s+1 s x31 x2 x32 −4 x24 −2 For s > 2, denote by D(s) the set of all the following monomials: ON THE PETERSON HIT PROBLEM s s+1 x1 x22 x23 −3 x24 −2 , s s+1 x1 x22 −1 x23 x24 −4 , s+1 s x1 x32 x23 −2 x24 −4 , s s+1 x1 x32 x23 −2 x24 −4 , s s+1 x31 x22 −3 x23 x24 −4 , s s+1 x1 x22 x32 −1 x24 −4 , s s+1 x21 −1 x2 x23 x24 −4 , s s+1 x31 x2 x23 −4 x24 −2 , s s+1 x31 x2 x23 −2 x42 −4 , s+1 s x31 x52 x23 −6 x42 −4 47 s+1 s x1 x22 x32 −4 x24 −1 , s s+1 x1 x32 x23 −4 x24 −2 , s+1 s x31 x2 x23 −2 x24 −4 , s+1 s x31 x22 −3 x23 x42 −4 , Set E(2) = C(2) ∪ {x31 x42 x3 x4 }, E(3) = C(3) ∪ D(3) ∪ {x31 x52 x63 x84 } and E(s) = s s+1 C(s) ∪ D(s) ∪ {x31 x52 x23 −6 x24 −4 }, for s > Proposition 5.5.1 For any integer s > 1, E(s) ∪ Φ0 (B3 (n)) ∪ ψ(B4 (m)) is the set of all the admissible monomials for A-module P4 in degree n = 2m + with m = 2s + 2s−1 − Lemma 5.5.2 Let x be an admissible monomial of degree n = 2s+t + 2s − in P4 If [x] ∈ KerSq ∗ , then either ω(x) = (2(s) , 1) Proof We prove the lemma by induction on s Since n = 2s+1 + 2s − is even, we get either ω1 (x) = or ω1 (x) = or ω1 (x) = If ω1 (x) = 0, then x = Sq (y) for some monomial y If ω1 (x) = 4, then x = X∅ y for some monomial y Since x is admissible, y also is admissible This implies KerSq ∗ ([x]) = [y] = and we have a contradiction So ω1 (x) = and x = xi xj y with i < j 4, and y a monomial of degree 2s + 2s−1 − in P4 Using Proposition 2.10 we get ωi (x) = s for i s Then x = x z with x , z monomials in P4 and deg z = 2t − By a direct computation we see that if w is a monomial such that either ω(w) = (2, 1, 3) or ω(w) = (2, 2, 3) or ω(w) = (2, 3, 2, 2) then w is strictly inadmissible Now, the lemma follows from this fact, Lemma 5.3.1 and Theorem 2.9 The following is proved by a direct computation Lemma 5.5.3 The following monomials are strictly inadmissible: i) x2i xj xm , x3i x4j x3m , x7i x7j x8m , i < j < m 4 10 8 3 12 3 12 ii) x1 x72 x10 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 , 8 x1 x2 x3 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 Proof of Proposition 5.5.1 Let x be an admissible monomial of degree n = 2s+1 + 2s − in P4 and [x] ∈ KerSq ∗ By Lemma 5.5.2, ωi (x) = 2, for i s, ωs+1 (x) = and ωi (x) = for i > s + By induction on s, we see that if x ∈ / E(s)∪Φ0 (B3 (n)) then there is a monomial w which is given in one of Lemmas 5.2.3, u 5.5.3 such that x = wy for some monomial y and positive integer u By Theorem 2.9, x is inadmissible Hence KerSq ∗ is spanned by the set [E(s) ∪ Φ0 (B3 (n))] in degree n = 2s+1 + 2s − Now, we prove that set [E(s) ∪ Φ0 (B3 (n))] is linearly independent It suffices to prove that the set [E(s)] is linearly independent For s = 2, |E(2)| = 12 Suppose there is a linear relation 12 S= γi di ≡ 0, (5.1) i=1 with γi ∈ F2 and di = d10,i A direct computation from the relations p(1;j) (S) ≡ 0, for j = 1, 2, 3, we obtain γi = for all i 48 NGUYỄN SUM For s > 2, |E(s)| = 26 Suppose there is a linear relation 26 S= γi di ≡ 0, (5.2) i=1 with γi ∈ F2 and di = dn,i A direct computation from the relations p(r;j) (S) ≡ 0, for r < j 4, we obtain γi = for all i The proposition is proved 5.6 The case of degree 2s+t+u + 2s+t + 2s − First, we determine the ω-vector of an admissible monomial of degree n = 2s+t+u + 2s+t + 2s − Lemma 5.6.1 If x is an admissible monomial of degree 2s+t+u + 2s+t + 2s − in P4 then ω(x) = (3(s) , 2(t) , 1(u) ) s+t+u s+t s −1 Proof Observe that z = x21 x2 −1 x23 −1 is the minimal spike of degree s+t+u s+t s +2 + − and ω(z) = (3(s) , 2(t) , 1(u) ) Since 2s+t+u + 2s+t + 2s − is odd and x is admissible, using Proposition 2.10 and Theorem 2.12, we get ωi (x) = s i−1 for i s Set x = i s XI2i−1 (x) Then x = x y for some monomial y We have ωj (y) = ωj+s (x) for all j and 2s+t+u + 2s+t + 2s − = deg x = 2i−1 ωi (x) i s = 3(2 − 1) + 2s 2j−1 ωj+s (x) j = 3.2s − + 2s deg y This equality implies deg y = 2t+u + 2u − Since x is admissible, using Theorem 2.9, we see that y is also admissible By a direct computation we see that if w is a monomial such that ω(w) = (3, 2, 3) then w is strictly inadmissible Combining this fact, Lemma 5.3.1, Proposition 2.10 and Theorem 2.9, we obtain ω(y) = (2(t) , 1(u) ) The lemma is proved Applying Theorem 1.3, we get the following Proposition 5.6.2 Let s, t, u be positive integers If s 3, then Φ(B3 (n)) is a minimal set of generators for A-module P4 in degree n = 2s+t+u + 2s+t + 2s − So, we need only to consider the cases s = and s = 5.6.1 The subcase s = t = For s = 1, t = 1, we have n = 2u+2 + According to Theorem 4.3, we have B3 (n) = ψ(Φ(B2 (2u+1 ))), ψ(Φ(B2 (8)) ∪ {x71 x92 x33 }, if u = 2, if u = Proposition 5.6.3 i) Φ(B3 (11)) ∪ {x31 x42 x3 x34 , x31 x42 x33 x4 } is the set of all the admissible monomials for A-module P4 in degree 11 3 12 3 11 11 ii) Φ(B3 (19)) ∪ {x71 x92 x23 x4 , x31 x12 x3 x4 , x x2 x3 x4 , x x2 x3 x4 , x x2 x3 x4 , 8 8 3 9 x1 x2 x3 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 , x1 x2 x3 x4 } is the set of all the admissible monomials for A-module P4 in degree 19 ON THE PETERSON HIT PROBLEM u+2 u+2 49 u+2 iii) Φ(B3 (n)) ∪ {x31 x42 x3 x42 −5 , x31 x42 x32 −5 x4 , x31 x42 x33 x42 −7 } is the set of all the admissible monomials for A-module P4 in degree n = 2u+2 +3, with any positive integer u By a direct computation, we can easy obtain the following lemma Lemma 5.6.4 The following monomials are strictly inadmissible: i) x31 x42 x43 x44 xi x3j , i, j > 1, i = j, x71 x32 x43 x44 xj , x31 x52 x53 x54 xj , j = 3, 4 18 10 16 2 24 2 16 ii) X2 x21 x2j x28 , Xj x1 x4 x2 x3 , X2 x1 xj x2 x3 x4 , Xj x1 x2 x3 x4 , Xj x1 x2 x3 x4 , 2 24 2 24 2 16 Xj x1 x2 xi x3 x4 , X3 x1 x2 xi x3 , X2 x1 x4 x2 x3 , i = 1, 2, j = 3, Proof of Theorem 5.6.3 Let x be an admissible monomial of degree n = 2u+2 + in P4 By Lemma 5.6.1, ω1 (x) = So x = Xi y with y a monomial of degree 2u+1 Since x is admissible, by Theorem 2.9, y ∈ B4 (2u+1 ) By a direct computation, we see that if x = Xi y with y ∈ B4 (2u+1 ) and x not belongs to the set C4 (n) as given in the proposition, then there is a monomial w which is given in one of r Lemmas 5.3.3, 5.6.4 such that x = wy for some monomial y and integer r > By Theorem 2.9, x is inadmissible Hence (QP4 )n is spanned by the set [C4 (n)] Set |C4 (2u+2 + 3) ∩ P4+ | = m(u), where m(1) = 32, m(2) = 80, m(u) = 64 for all u > Suppose that there is a linear relation m(u) S= γi di = 0, (5.6.1) i=1 with γi ∈ F2 and di = dn,i By a direct computation from the relations p(j;J) (S) ≡ with (j; J) ∈ N4 , we obtain γi = for all i if u = For u = 2, γj = for j =1, 3, 4, 6, 7, 8, 9, 10, 11, 12, 14, 16, 17, 18, 19, 21, 23, 26, 27, 28, 29, 30, 31, 32, 35, 36, 38, 40, 43, 45, 51, 54, 55, 60, 61, 62, 68, 71, 79, 80, and γ2 = γi , i = 5, 24, 25, 41, 42, 52, 53, γ13 = γi , i = 13, 33, 20, 56, 48, 58, γ15 = γi , i = 22, 34, 49, 57, 59, γ37 = γi , i = 67, 70, 75, γ46 = γi , i = 69, 72, 76, γ65 = γi , i = 66, 73, 74, 77, 78, γ46 = γ39 + γ2 , γ44 = γ37 + γ2 , γ65 = γ47 + γ13 , γ65 = γ50 + γ22 , γ63 = γ37 + γ13 , γ64 = γ46 + γ22 Substituting the above equalities into the relation (5.6.1), we have γ37 [θ1 ] + γ46 [θ2 ] + γ13 [θ3 ] + γ22 [θ4 ] + γ65 [θ5 ] + γ2 [θ6 ] = 0, (5.6.2) where θ1 = d37 + d44 + d63 + d67 + d70 + d75 , θ2 = d39 + d46 + d64 + d69 + d72 + d76 , θ3 = d13 + d20 + d33 + d47 + d48 + d56 + d58 + d63 , θ4 = d15 + d22 + d34 + d49 + d50 + d57 + d59 + d64 , θ5 = d47 + d50 + d65 + d66 + d73 + d74 + d77 + d78 , θ6 = d2 + d5 + d24 + d25 + d39 + d41 + d42 + d44 + d52 + d53 We need to prove γ2 = γ13 = γ22 = γ37 = γ46 = γ65 = The proof is divided into steps Step First we prove γ65 = by showing the polynomial [θ] = [β1 θ1 + β2 θ2 + β3 θ3 + β4 θ4 + θ5 + β6 θ6 ] = for all β1 , β2 , β3 , β4 , β6 ∈ F2 Suppose the contrary that this polynomial is hit Then we have θ = Sq (A) + Sq (B) + Sq (C) + Sq (D), 50 NGUYỄN SUM for some polynomials A, B, C, D in P4+ Let (Sq )3 act on the both sides of this equality Using the relations (Sq )3 Sq = 0, (Sq )3 Sq = 0, we get (Sq )3 (θ) = (Sq )3 Sq (C) + (Sq )3 Sq (D) 2 12 The monomial x71 x12 x3 x4 is a term of (Sq ) (θ) If x1 x2 x3 x4 is a term of the poly2 nomial (Sq ) Sq (y) with y a monomial of degree 11 in P4 , then y = x71 f1 (z) with z a monomial of degree in P3 Then x71 x12 x3 x4 is a term of x1 (Sq ) Sq (f1 (z)) = 12 2 This is a contradiction So x1 x2 x3 x4 is not a term of (Sq ) Sq (D) for all D 2 7 Hence x71 x12 x3 x4 is a term of (Sq ) Sq (C), then either x1 x2 x3 x4 or x1 x2 x3 x4 or x1 x2 x3 x4 is a term of C Suppose x71 x52 x23 x4 is a term of C Then (Sq )3 (θ + Sq (x71 x52 x23 x4 )) = (Sq )3 (Sq (C ) + Sq (D)), where C = C + x71 x52 x23 x4 We see that the monomial x16 x2 x3 x4 is a term of (Sq ) (θ + Sq (x1 x2 x3 x4 )) This monomial is not a term of (Sq )3 Sq (D) for all D So it is a term of (Sq )3 Sq (C ) Then either x71 x52 x23 x4 or x71 x62 x3 x4 is a term of C Since x71 x52 x23 x4 is a term of C , x71 x62 x3 x4 is s term of C Hence we obtain (Sq )3 (θ + Sq (x71 x52 x23 x4 + x71 x62 x3 x4 )) = (Sq )3 (Sq (C ) + Sq (D)), where C = C + x71 x52 x23 x4 + x71 x62 x3 x4 Now x71 x12 x3 x4 is a term of (Sq )3 (θ + Sq (x71 x52 x23 x4 + x71 x62 x3 x4 )) So either x71 x52 x3 x24 or x71 x52 x23 x4 or x71 x62 x3 x4 is a term of C Since x71 x52 x23 x4 + x71 x62 x3 x4 is a summand of C , x71 x52 x3 x24 is s term of C Then x16 x2 x3 x4 is a 7 7 term of (Sq ) (θ + Sq (x1 x2 x3 x4 + x1 x2 x3 x4 + x1 x2 x3 x4 )) So either x1 x2 x3 x24 or x71 x52 x23 x4 or x71 x62 x3 x4 is a term of C + x71 x52 x3 x24 and we have a contradiction By a same argument, if either x71 x52 x3 x24 or x71 x62 x3 x4 is a term of C then we have also a contradiction Hence [θ] = and γ65 = Step By a direct computation, we see that the homomorphism ϕ3 sends (5.6.2) to γ37 [θ1 ] + γ2 [θ3 ] + γ22 [θ4 ] + γ46 [θ5 ] + γ13 [θ6 ] = By Step 1, we obtain γ46 = Step The homomorphism ϕ2 sends (5.6.2) to γ13 [θ1 ] + γ22 [θ2 ] + γ37 [θ3 ] + γ2 [θ6 ] = By Step 2, we obtain γ22 = Step Now the homomorphism ϕ3 sends (5.6.2) to γ37 [θ2 ] + γ13 [θ4 ] + γ2 [θ6 ] = Combining Step and Step 3, we obtain γ13 = γ37 = Since ϕ2 ([θ3 ]) = [θ6 ], we get γ2 = So we obtain γj = for all j The proposition follows 5.6.2 The subcase s = 1, t = For s = 1, t = 2, we have n = 2u+3 + = 2m + with m = 2u+2 + Combining Theorem 1.3 and Theorem 4.3, we have B3 (n) = ψ(Φ(B2 (m))) where B2 (m) = {x31 x72 , x71 x32 }, u+2 u+2 u+2 {x31 x22 −1 , x12 −1 x32 , x71 x22 −5 }, if u = 1, if u > ON THE PETERSON HIT PROBLEM 51 Denote by F (u) the set of all the following monomials: x31 x42 x3 x24 x31 x72 x23 u+3 u+3 , x31 x42 x32 u+3 −1 x4 , x31 x22 u+3 −1 x3 x4 , x21 x4 , x71 x32 x32 u+3 −4 x4 , x71 x22 u+3 −5 x3 x4 , x71 x72 x32 −1 −4 u+3 x31 x42 x33 x24 −3 , x71 x32 x43 x24 u+3 −7 u+3 x31 x42 x32 −5 x54 , , x31 x72 x83 x42 u+3 −11 u+3 x31 x42 x73 x42 −7 , , x71 x32 x83 x42 u+3 −11 u+3 −1 x2 x3 x4 , u+3 −8 x4 , u+3 x31 x72 x43 x42 −7 , Proposition 5.6.5 i) Φ(B3 (23)) ∪ F (1) ∪ {x71 x92 x23 x54 , x71 x92 x33 x44 } is the set of all the admissible monomials for A-module P4 in degree 23 u+3 u+3 2u+3 −11 ii) Φ(B3 (n)) ∪ F (u) ∪ {x71 x72 x83 x42 −15 , x71 x72 x93 x42 −16 , x31 x42 x11 } is the x4 set of of all the admissible monomials for A-module P4 in degree n = 2u+3 + with any positive integer u > By a direct computation, we can easy obtain the following lemma Lemma 5.6.6 The following monomials are strictly inadmissible: 8 4 i) X2 x21 x6j x12 , Xj x1 x2 x3 x4 , X2 x1 xi x2 x3 x4 , X2 x1 x2 x3 x4 , i = 1, 2, j = 3, 20 2 20 2 12 16 14 16 ii) X3 x21 x22 x12 i x3 , X3 x1 x2 xi x3 x4 , Xj x1 x2 xi x3 x4 , Xj x1 x2 xi x3 , 10 16 10 16 10 20 14 16 Xj x1 x2 x3 x4 , Xj x1 x2 x3 x4 , X3 x1 x2 x3 , X2 x1 x2 x3 x4 , i = 1, 2, j = 3, Proof of Proposition 5.6.5 Let x be an admissible monomial of degree n = 2u+3 +7 in P4 By Lemma 5.6.1, ω1 (x) = So x = Xi y with y a monomial of degree 2u+2 + Since x is admissible, by Theorem 2.9, y ∈ B4 (2u+2 + 2) By a direct computation, we see that if x = Xi y with y ∈ B4 (2u+2 + 2) and x not belongs to the set C4 (n) as given in the proposition, then there is a monomial w r which is given in one of Lemmas 5.6.6, 5.3.3 such that x = wy for some monomial y and integer r > By Theorem 2.9, x is inadmissible Hence (QP4 )n is spanned by the set [C4 (n)] For u = 1, we have, |C4+ (23) ∩ P4+ | = 99 Suppose that there is a linear relation 99 S= γi di = 0, (5.6.1) i=1 with γi ∈ F2 and di = d23,i By a direct computation from the relations p(j;J) (S) ≡ with (j; J) ∈ N4 , we obtain γi = for all i ∈ E, with some E ⊂ N99 and the relation (5.6.2) becomes 15 ci [θi ] = 0, i=1 (5.6.2) 52 NGUYỄN SUM where c1 = γ1 , c2 = γ4 , c3 = γ33 , c4 = γ94 , c5 = γ2 , c6 = γ22 , c7 = γ74 , c8 = γ29 , c9 = γ81 , c10 = γ68 , c11 = γ10 , c12 = γ43 , c13 = γ54 , c14 = γ70 , c15 = γ11 and θ1 = d1 + d17 + d37 + d49 , θ2 = d4 + d21 + d44 + d53 , θ3 = d33 + d36 + d72 + d73 , θ4 = d94 + d97 + d98 + d99 , θ5 = d2 + d19 + d40 + d51 , θ6 = d22 + d25 + d62 + d63 , θ7 = d74 + d77 + d82 + d83 , θ8 = d12 + d14 + d26 + d29 + d66 + d67 , θ9 = d40 + d42 + d78 + d81 + d86 + d87 , θ10 = d10 + d15 + d24 + d27 + d46 + d47 + d64 + d65 , θ11 = d38 + d43 + d46 + d47 + d76 + d79 + d84 + d85 , θ12 = d62 + d67 + d68 + d71 + d88 + d89 + d92 + d93 , θ13 = d47 + d54 + d57 + d62 + d69 + d82 + d85 + d88 + d90 , θ14 = d12 + d15 + d19 + d20 + d46 + d47 + d51 + d52 + d58 + d61 + d64 + d66 + d67 + d70 + d84 + d87 + d89 + d91 , θ15 = d11 + d12 + d18 + d20 + d24 + d25 + d26 + d27 + d38 + d40 + d45 + d47 + d48 + d50 + d52 + d57 + d61 + d63 + d64 + d65 + d66 + d67 + d69 + d77 + d78 + d83 + d85 + d86 + d87 + d89 + d90 Now, we show that ci = for i = 1, 2, , 15 The proof is divided into steps 15 Step Set θ = θ1 + i=2 βi θi for βi ∈ F2 , i = 2, 3, , 15 We prove that [θ] = Suppose the contrary that θ is hit Then we have θ = Sq (A) + Sq (B) + Sq (C) + Sq (D) for some polynomials A, B, C, D ∈ P4+ Let (Sq )3 act to the both sides of the above equality, we obtain (Sq )3 (θ) = (Sq )3 Sq (C) + (Sq )3 Sq (D) By a similar computation as in the proof of Proposition 5.4.5, we see that the monomial x81 x42 x23 x15 is a term of (Sq ) (θ) This monomial is not a term of (Sq ) (Sq (C) + Sq (D)) for all polynomials C, D and we have a contradiction So [θ] = and we get c1 = γ1 = By an argument analogous to the previous one, we get c2 = c3 = c4 = Now, the relation (5.6.2) becomes 15 ci [θi ] = i=5 Step The homomorphisms ϕ1 , ϕ1 ϕ3 , ϕ1 ϕ3 ϕ4 , ϕ1 ϕ3 ϕ2 , ϕ1 ϕ3 ϕ2 ϕ4 , ϕ1 ϕ3 ϕ4 ϕ2 ϕ3 (5.6.3) ON THE PETERSON HIT PROBLEM 53 send (5.6.3) respectively to c10 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c9 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c7 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c8 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c6 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c5 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] Using the results in Step 1, we get c5 = c6 = c7 = c8 = c9 = c10 = So the relation (5.6.3) becomes c11 [θ11 ] + c12 [θ12 ] + c13 [θ13 ] + c14 [θ14 ] + c15 [θ15 ] = (5.6.4) Step The homomorphism ϕ1 sends (5.6.4) to c13 [θ6 ] + (c14 + c15 )[θ7 ] + (c11 + c12 )[θ11 ] + c12 [θ12 ] + c13 [θ13 ] + c14 [θ14 ] + c15 [θ15 ] = By Step 2, we get c13 = and c14 = c15 So the relation (5.6.4) becomes c11 [θ11 ] + c12 [θ12 ] + c14 [θ14 ] + c14 [θ15 ] = (5.6.5) Step The homomorphism ϕ3 sends (5.6.5) to c11 [θ11 ] + c14 [θ12 ] + (c12 + c14 )[θ13 ] + c14 [θ14 ] + c14 [θ15 ] = By Step 3, we get c12 = c14 Then the relation (5.6.5) becomes c11 [θ11 ] + c12 [θ12 ] + c12 [θ14 ] + c12 [θ15 ] = (5.6.6) Step The homomorphism ϕ2 sends (5.6.6) to (c11 + c12 )[θ12 ] + c12 [θ14 ] + c12 [θ15 ] = From the result in Step 4, we get c11 = Then the relation (5.6.6) becomes c12 ([θ12 ] + [θ14 ] + [θ15 ]) = (5.6.7) Step The homomorphism ϕ1 sends (5.6.7) to c12 [θ11 ] + c12 ([θ12 ] + [θ14 ] + [θ15 ]) = By the result in Step 5, we have c12 = The case u = of the proposition is completely proved For u > 1, we have |C4 (n) ∩ P4+ | = 141 Suppose that there is a linear relation 141 S= γi di = 0, (5.6.8) i=1 with γi ∈ F2 and di = dn,i ∈ B4+ (n) By a direct computation from the relations p(j;J) (S) ≡ with (j; J) ∈ N4 , we obtain γi = for all i ∈ / E, with some E =⊂ N141 and the relation (5.6.8) becomes 15 ci [θi ] = 0, i=1 (5.6.9) 54 NGUYỄN SUM where c1 = γ1 , c2 = γ6 , c3 = γ51 , c4 = γ136 , c5 = γ2 , c6 = γ31 , c7 = γ107 , c8 = γ40 , c9 = γ116 , c10 = γ101 , c11 = γ14 , c12 = γ56 , c13 = γ79 , c14 = γ23 , c15 = γ15 and θ1 = d1 + d25 + d55 + d73 , θ2 = d6 + d30 + d66 + d78 , θ3 = d51 + d54 + d105 + d106 , θ4 = d7 + d8 + d47 + d48 , θ5 = d2 + d27 + d58 + d75 , θ6 = d31 + d34 + d89 + d90 , θ7 = +d107 + d110 + d117 + d118 , θ8 = d16 + d22 + d35 + d40 + d94 + d95 , θ9 = d58 + d64 + d111 + d116 + d122 + d123 , θ10 = d89 + d95 + d101 + d104 + d124 + d127 + d129 + d130 , θ11 = d14 + d19 + d33 + d36 + d68 + d69 + d91 + d92 , θ12 = d56 + d61 + d68 + d69 + d109 + d112 + d119 + d120 , θ13 = d67 + d69 + d79 + d82 + d89 + d90 + d117 + d118 + d124 + d125 , θ14 = d16 + d23 + d27 + d29 + d70 + d71 + d72 + d75 + d77 + d83 + d88 + d94 + d95 + d122 + d123 + d126 + d127 , θ15 = d15 + d19 + d26 + d27 + d33 + d34 + d35 + d36 + d58 + d61 + d68 + d69 + d70 + d74 + d75 + d82 + d83 + d91 + d92 + d109 + d110 + d111 + d112 + d119 + d120 + d125 Now, we prove ci = for i = 1, 2, , 15 The proof is divided into steps 15 Step First, we prove c1 = Set θ = θ1 + j=2 cj θj We show that [θ] = for all cj ∈ F2 , j = 2, 3, , 15 Suppose the contrary that θ is hit Then we have u+2 m Sq (Am ), θ= m=0 for some polynomials Am , m = 0, 1, , u + Let (Sq )3 act on the both sides of this equality Since (Sq )3 Sq = 0, (Sq )3 Sq = 0, we get u+2 m (Sq )3 (θ) = (Sq )3 Sq (Am ) m=2 u+3 It is easy to see that the monomial x = x81 x42 x23 x42 −1 is a term of (Sq )3 (θ), hence m it is a term of (Sq )3 Sq (y) for some monomial y of degree 2u+3 − 2m + with u+3 m Then y = x22 −1 f2 (z) with z a monomial of degree − 2m in P3 and u+3 m m x is a term of x22 −1 (Sq )3 Sq (z) If m > then Sq (z) = If m = the m Sq (z) = z , hence (Sq )3 Sq (z) = (Sq )3 (z ) = So x is not a term of u+2 m (Sq )3 (θ) = (Sq )3 Sq (Am ), m=2 for all polynomial Am with m > This is a contradiction So we get c1 = ON THE PETERSON HIT PROBLEM 55 By an argument analogous to the previous one, we get c2 = c3 = c4 = Then the relation (5.6.9) becomes 15 ci [θi ] = (5.6.10) i=5 Step The homomorphisms ϕ1 , ϕ1 ϕ3 , ϕ1 ϕ3 ϕ4 , ϕ1 ϕ3 ϕ2 , ϕ1 ϕ3 ϕ2 ϕ4 , ϕ1 ϕ3 ϕ4 ϕ2 ϕ3 send (5.6.3) respectively to c10 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c9 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c7 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c8 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c6 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] , c5 [θ3 ] = mod [θ5 ], [θ6 ], , [θ15 ] By Step 1, we get c5 = c6 = c7 = c8 = c9 = c10 = So the relation (5.6.3) becomes c11 [θ11 ] + c12 [θ12 ] + c13 [θ13 ] + c14 [θ14 ] + c15 [θ15 ] = (5.6.11) Step Applying the homomorphism ϕ1 to (5.6.11), we get c13 [θ6 ] + c14 [θ8 ] + (c11 + c12 + c15 )[θ11 ] + c12 [θ12 ] + c13 [θ13 ] + c14 [θ14 ] + c15 [θ15 ] = By the results in Step 2, we obtain c13 = c14 = Then the relation (5.6.11) becomes c11 [θ11 ] + c12 [θ12 ] + c14 [θ15 ] = (5.6.12) Step Applying the homomorphism ϕ3 to the relation (5.6.12) we obtain c11 [θ11 ] + c12 [θ13 ] + c15 [θ15 ] = By the results in Step 3, we get c12 = So the relation (5.6.12) becomes c11 [θ11 ] + c15 [θ15 ] = (5.6.13) Step Applying the homomorphism ϕ2 to the relation (5.6.12) one gets c11 [θ13 ] + c15 [θ15 ] = By Step 4, we get c10 = γ41 = So the relation 5.6.13 becomes c15 [θ15 ] = (5.6.14) Step Applying the homomorphism ϕ1 to the relation 5.6.14 we obtain c15 [θ11 ] + c15 [θ15 ] = By Step 5, we get c15 The proposition is completely proved 5.6.3 The subcase s = 1, t > For s = 1, t > 2, we have n = 2t+u+1 + 2t+1 − = 2m + with m = 2t+u + 2t − From Theorem 4.3, we have B3 (n) = ψ(Φ(B2 (m))) 56 NGUYỄN SUM Proposition 5.6.7 t+1 t+2 t+2 t+1 i) Φ(B3 (n)) ∪ {x31 x42 x23 −5 x42 −3 , x31 x42 x32 −5 x42 −3 } is the set of of all the admissible monomials for A-module P4 in degree n = 2t+2 + 2t+1 − with any positive integer t > ii) Φ(B3 (n)) ∪ A(t, u) is the set of of all the admissible monomials for A-module P4 in degree n = 2t+u+1 + 2t+1 − with any positive integers t > 2, u > 1, where A(t, u) is the set consisting of monomials: x31 x42 x23 t+1 −5 2t+u+1 −3 x4 , x31 x42 x32 t+u+1 −5 2t+1 −3 x4 , x31 x42 x32 t+2 −5 2t+u+1 −2t+1 −3 x4 By a direct computation, we can easy obtain the following lemma Lemma 5.6.8 The following monomials are strictly inadmissible: 2 12 10 12 16 X3 x21 x22 x83 x28 xi , X3 x1 x2 x3 x4 xi , i = 1, 2, X4 x1 x2 x3 x4 Proof of Proposition 5.6.7 Let x ∈ P4 be an admissible monomial of degree n = 2t+u+1 + 2t+1 − By Lemma 5.6.1, ω1 (x) = So x = Xi y with y a monomial of degree 2t+u + t − Since x is admissible, by Theorem 2.9, y ∈ B4 (2t+u + 2t − 2) By a direct computation, we see that if x = Xi y with y ∈ B4 (2t+u + 2t − 2) and x not belongs to the set C4 (n) as given in the proposition, then there is a monomial r w which is given in one of Lemmas 5.6.8 and 5.3.3 such that x = wy for some monomial y and integer r > By Theorem 2.9, x is inadmissible Hence (QP4 )n is spanned by the set [C4 (n)] We set |C4 (n) ∩ P4+ | = m(t, u) with m(t, 1) = 84 for u = and m(t, u) = 126 for u > Suppose that there is a linear relation m(t,u) S= γi di = 0, i=1 with γi ∈ F2 and di = dn,i By a direct computation from the relations p(j;J) (S) ≡ with (j; J) ∈ N4 , we obtain γi = for all i 5.6.4 The subcase s = 2, t = For s = 2, t = 1, we have n = 2u+3 + According to Theorem 4.3, we have B3 (n) = ψ (Φ(B2 (2u+1 ))), 19 ψ (Φ(B2 (8))) ∪ {x15 x2 x3 }, if u = 2, if u = Denote by G(u) the set of monomials: x31 x72 x23 u+3 x31 x72 x73 x24 −5 x4 , u+3 −8 x71 x32 x32 u+3 , x71 x32 x73 x42 −5 x4 , u+3 −8 x71 x22 u+3 −5 x3 x4 , , x71 x72 x33 x42 u+3 −8 , x71 x72 x23 u+3 −8 x4 , Proposition 5.6.9 i) Φ(B3 (25)) ∪ G(1) ∪ {x71 x92 x33 x64 } is the set of of all the admissible monomials for A-module P4 in degree 25 ii) Φ(B3 (n)) ∪ G(u) ∪ H(u) is the set of of all the admissible monomials for A-module P4 in degree n = 2u+3 + with any positive integer u > 1, where H(u) ON THE PETERSON HIT PROBLEM 57 is the set consisting of monomials: x31 x72 x11 x4 x71 x72 x83 x24 u+3 u+3 −12 −13 , x71 x32 x11 x4 , x71 x72 x11 x4 u+3 u+3 −12 −16 , x71 x11 x3 x4 u+3 −12 , The following is proved by a direct computation Lemma 5.6.10 The following monomials are strictly inadmissible: i) X3 X22 x41 x82 x44 , Xj X22 x41 x82 x44 , X33 x4i x83 x44 , X23 x41 x82 x4j , i = 1, 2, j = 3, 12 16 4 16 16 4 24 12 16 ii) X4 X32 x12 x2 x3 , X4 X2 x1 x2 x4 , X4 xi x3 x4 , X4 X2 x1 x2 x4 , X4 X3 x1 x2 xi x3 , 12 16 12 16 12 16 4 20 4 16 Xj X2 x1 x2 x3 , Xj X2 x1 x2 x4 , X4 X2 x1 x2 x4 , Xj x1 x2 xi xj , X2 x1 x2 xj , 16 4 16 12 16 16 12 16 X43 x4i x12 x4 , X4 xi x3 x4 , X3 xi x3 x4 , Xj x1 x2 x3 x4 , X4 X2 x1 x2 x3 x4 16 X4 x1 x2 x3 x4 , i = 1, 2, j = 3, Proof of Proposition 5.6.9 Let x be an admissible monomial of degree n = 2u+3 +9 in P4 By Lemma 5.6.1, ω1 (x) = ω2 (x) = So x = Xi Xj2 y with y a monomial of degree 2u+1 Since x is admissible, by Theorem 2.9, y ∈ B4 (2t+u + 2t − 2) By a direct computation, we see that if x = Xi Xj2 y with y ∈ B4 (2t+u + 2t − 2) and x not belongs to the set C4 (n) given in the proposition, then there is a monomial r w which is given in one of Lemmas 5.6.10, 5.3.3 such that x = wy for some monomial y and integer r > By Theorem 2.9, x is inadmissible Hence (QP4 )n is spanned by the set [C4 (n)] We denote |C4 (n) ∩ P4+ | = m(u) with m(1) = 88, m(2) = 165 and m(u) = 154 for u Suppose that there is a linear relation m(u) S= γi di = 0, i=1 with γi ∈ F2 and di = dn,i By a direct computation from the relations p(j;J) (S) ≡ with (j; J) ∈ N4 , we obtain γi = for all i 5.6.5 The subcase s = 2, t For s = 2, t 2, we have n = 2t+u+2 + 2t+2 + = 4m + with m = 2t+u + 2t − From Theorem 1.3, we have B3 (n) = ψ (Φ(B2 (m))) Denote by B(t, u) the set of monomials: x31 x72 x23 t+2 −5 2t+u+2 −4 x4 , x71 x32 x23 t+2 x31 x72 x23 t+u+2 x71 x72 x23 t+2 −5 2t+u+2 −4 x4 , −5 2t+2 −4 x4 , x71 x32 x32 t+u+2 −5 2t+2 −4 x4 , −8 2t+u+2 −5 x4 , x71 x72 x32 t+u+2 −8 2t+2 −5 x4 , x71 x22 t+2 −5 2t+u+2 −4 x3 x4 , x71 x22 t+u+2 −5 2t+2 −4 x3 x4 , and by C(t, u) the set of monomials: x31 x72 x23 x71 x22 t+3 t+3 −5 2t+u+2 −2t+2 −4 x4 , x71 x32 x32 t+3 −5 2t+u+2 −2t+2 −4 x4 , −5 2t+u+2 −2t+2 −4 x3 x4 , x71 x72 x32 t+3 −8 2t+u+2 −2t+2 −5 x4 58 NGUYỄN SUM Proposition 5.6.11 i) Φ(B3 (n)) ∪ B(t, 1) is the set of all the admissible monomials for A-module P4 in degree n = 2t+3 + 2t+2 + ii) For any positive integer t, u > 1, Φ(B3 (n)) ∪ B(t, u) ∪ C(t, u) is the set of all the admissible monomials for A-module P4 in degree n = 2t+u+2 + 2t+2 + By a direct computation, we get the following Lemma 5.6.12 The following monomials are strictly inadmissible: 12 16 12 12 16 12 12 16 4 8 16 12 16 Xj X32 x12 x2 x3 , X4 xi x3 x4 , X4 x1 x2 x4 , X4 x1 x2 x3 x4 xj , X4 X3 x3 x1 x4 x2 , 4 8 16 4 8 16 4 8 16 X4 X3 x1 x2 x4 xi x3 , Xj x1 x2 x3 xi x4 , X4 x1 x3 x2 x3 x4 , i = 1, 2, j = 3, Proof of Proposition 5.6.11 Let x ∈ P4 be an admissible monomial of degree n = 2t+u+2 + 2t+2 + By Lemma 5.6.1, ω1 (x) = ω2 (x) = So x = Xi Xj2 y with y a monomial of degree 2t+u + 2t − Since x is admissible, by Theorem 2.9, y ∈ B4 (2t+u + 2t − 2) By a direct computation, we see that if x = Xi Xj2 y with y ∈ B4 (2t+u + 2t − 2) and x not belongs to the set C4 (n) as given in the proposition, then there is a monomial w r which is given in one of Lemmas 5.6.12, 5.1.3 such that x = wy for some monomial y and integer r > By Theorem 2.9, x is inadmissible Hence (QP4 )n is spanned by the set [C4 (n)] We set |C4 (n) ∩ P4+ | = m(t, u) with m(t, 1) = 154 and m(t, u) = 231 for t Suppose that there is a linear relation m(t,u) S= γi di = 0, i=1 with γi ∈ F2 and di = dn,i By a direct computation from the relations p(j;J) (S) ≡ with (j; J) ∈ N4 , we obtain γi = for all i Acknowledgment I would like to thank Prof Nguyễn H V Hưng for helpful suggestions and constant encouragement My thanks also go to all colleagues at the Department of Mathematics, Quy Nhơn University for many conversations The final version of this work was completed while the author was visiting the Vietnam Institute for Advanced Study in Mathematics (VIASM) He would like to thank the VIASM for supporting the visit and hospitality The work was also supported in part by the Project Grant No B2013.28.129 References [1] J M Boardman, Modular representations on the homology of powers of real projective space, in: M.C Tangora (Ed.), Algebraic Topology, Oaxtepec, 1991, in: Contemp Math., vol 146, 1993, pp 49-70, MR1224907 [2] R R Bruner, L M Hà and N H V Hưng, On the behavior of the algebraic transfer, Trans Amer Math Soc 357 (2005) 473-487, MR2095619 [3] D P Carlisle and R M W Wood, The boundedness conjecture for the action of the Steenrod algebra on polynomials, in: N Ray and G Walker (ed.), Adams Memorial Symposium on Algebraic Topology 2, (Manchester, 1990), in: London Math Soc Lecture Notes Ser., Cambridge Univ Press, Cambridge, vol 176, 1992, pp 203-216, MR1232207 [4] M C Crabb and J R Hubbuck, Representations of the homology of BV and the Steenrod algebra II, in: Algebraic Topology: New Trend in Localization and Periodicity, (Sant Feliu de Guớxols, 1994), in: Progr Math., Birkhă auser Verlag, Basel, Switzerland, vol 136, 1996, pp 143-154, MR1397726 ON THE PETERSON HIT PROBLEM 59 [5] V Giambalvo and F P Peterson, A-generators for ideals in the Dickson algebra, J Pure Appl Algebra 158 (2001), 161-182, MR1822839 [6] L M Hà, Sub-Hopf algebras of the Steenrod algebra and the Singer transfer, in: Proceedings of the International School and Conference in Algebraic Topology, Hà Nˆ ọi 2004, Geom Topol Monogr., Geom Topol Publ., Coventry, vol 11, 2007, 81-105, MR2402802 [7] N H V Hưng, The cohomology of the Steenrod algebra and representations of the general linear groups, Trans Amer Math Soc 357 (2005) 4065-4089, MR2159700 [8] N H V Hưng and T N Nam, The hit problem for the Dickson algebra, Trans Amer Math Soc 353 (2001) 5029-5040, MR1852092 [9] N H V Hưng and T N Nam, The hit problem for the modular invariants of linear groups, Journal of Algebra 246 (2001) 367-384, MR2872626 [10] N H V Hưng and F P Peterson, A-generator for the Dickson algebra, Trans Amer Math Soc., 347 (1995) 4687-4728, MR1316852 [11] N H V Hưng and F P Peterson, Spherical classes and the Dickson algebra, Math Proc Camb Phil Soc 124 (1998) 253–264, MR1631123 [12] A S Janfada and R M W Wood, The hit problem for symmetric polynomials over the Steenrod algebra, Math Proc Cambridge Philos Soc.133 (2002) 295-303, MR1912402 [13] A S Janfada and R M W Wood, Generating H ∗ (BO(3), F2 ) as a module over the Steenrod algebra, Math Proc Cambridge Philos Soc 134 (2003), 239-258, MR1972137 [14] M Kameko, Products of projective spaces as Steenrod modules, PhD Thesis, The Johns Hopkins University, ProQuest LLC, Ann Arbor, MI, 1990, 29 pp., MR2638633 [15] M Kameko, Generators of the cohomology of BV3 , J Math Kyoto Univ 38 (1998) 587-593, MR1661173 [16] M Kameko, Generators of the cohomology of BV4 , Toyama University, Preprint, 2003 [17] N Minami, The iterated transfer analogue of the new doomsday conjecture, Trans Amer Math Soc 351 (1999) 2325-2351, MR1443884 [18] M F Mothebe, Admissible monomials and generating sets for the polynomial algebra as a module over the Steenrod algebra, Afr Diaspora J Math 16 (2013), 18-27, MR3091712 [19] T N Nam, A-générateurs génériques pour l’algèbre polynomiale, Adv Math 186 (2004) 334-362, MR2073910 [20] T N Nam, Transfert algébrique et action du groupe linéaire sur les puissances divisées modulo 2, Ann Inst Fourier (Grenoble) 58 (2008), (5), 1785-1837, MR2445834 [21] F P Peterson, Generators of H ∗ (RP ∞ × RP ∞ ) as a module over the Steenrod algebra, Abstracts Amer Math Soc No 833 (April 1987) [22] F P Peterson, A-generators for certain polynomial algebras, Math Proc Cambridge Philos Soc 105 (1989), 311–312, MR0974987 [23] S Priddy, On characterizing summands in the classifying space of a group, I, Amer Jour Math 112 (1990) 737-748, MR1073007 [24] J Repka and P Selick, On the subalgebra of H∗ ((RP ∞ )n ; F2 ) annihilated by Steenrod operations, J Pure Appl Algebra 127 (1998) 273-288, MR1617199 [25] J H Silverman, Hit polynomials and the canonical antiautomorphism of the Steenrod algebra, Proc Amer Math Soc 123 (1995) 627-637, MR1254854 [26] J H Silverman and W M Singer, On the action of Steenrod squares on polynomial algebras II, J Pure Appl Algebra 98 (1995), 95-103, MR1317001 [27] W M Singer, The transfer in homological algebra, Math Zeit 202 (1989), 493-523, MR1022818 [28] W M Singer, On the action of the Steenrod squares on polynomial algebras, Proc Amer Math Soc 111 (1991) 577-583, MR1045150 [29] N E Steenrod and D B A Epstein, Cohomology operations, Ann of Math Stud vol 50, Princeton Univ Press, Princeton, N.J 1962, MR0145525 [30] N Sum, The hit problem for the polynomial algebra of four variables, Quy Nhơn University, Viˆ ẹt Nam, Preprint, 2007, 240 pp [31] N Sum, The negative answer to Kameko’s conjecture on the hit problem, C R Acad Sci Paris, Ser I 348 (2010), 669-672, MR2652495 [32] N Sum, The negative answer to Kameko’s conjecture on the hit problem, Adv Math 225 (2010), 2365-2390, MR2680169 60 NGUYỄN SUM [33] G Walker and R M W Wood, Young tableaux and the Steenrod algebra, in: Proceedings of the International School and Conference in Algebraic Topology, Hà Nˆ ọi 2004, Geom Topol Monogr., Geom Topol Publ., Coventry, vol 11, 2007, 379-397, MR2402815 [34] G Walker and R M W Wood, Weyl modules and the mod Steenrod algebra, J Algebra 311 (2007), 840-858, MR2314738 [35] G Walker and R M W Wood, Flag modules and the hit problem for the Steenrod algebra, Math Proc Cambridge Philos Soc 147 (2009), 143-171, MR2507313 [36] R M W Wood, Steenrod squares of polynomials and the Peterson conjecture, Math Proc Cambridge Phil Soc 105 (1989) 307-309, MR0974986 [37] R M W Wood, Problems in the Steenrod algebra, Bull London Math Soc 30 (1998) 449517, MR1643834 [38] R M W Wood, Hit problems and the Steenrod algebra, in: Proceedings of the summer school “Interactions between algebraic topology and invariant theory”, a satellite conference of the third European congress of mathematics, Lecture Course, Ioannina University, Greece, June 2000, Published in Ioannina University reports, 2001, pp 65-103 Department of Mathematics, Quy Nhơn University, - ịnh, Viˆẹt Nam 170 An Dương Vương, Quy Nhơn, Bình D E-mail: nguyensum@qnu.edu.vn ... Wood [37, 38] and others.) The µ-function is one of the numerical functions that have much been used in the context of the hit problem For a positive integer n, by µ(n) one means the smallest number... In Section 4, we recall the results on the hit problem for k Theorem 1.4 will be proved in Section by explicitly determining all of the admissible monomials in P4 The first formulation of this... reduced the length of the paper Preliminaries In this section, we recall some results in Kameko [14] and Singer [28] which will be used in the next sections ON THE PETERSON HIT PROBLEM Notation 2.1